Market Structure

The Times, They Are A-Changin’! Even for Treasuries Trading…

Recently, the Fed auctioned a trillion dollars of treasuries in a single month. CME acquired BrokerTec.  Primary dealers don’t have the balance sheets that they used to have.

What do they all have to do with one another?

Nope, not Bob Dylan.

  • The size of treasury markets is growing very rapidly, and so will treasury trading.

  • The old guard can’t keep up.

  • Microstructure changes are here and we’re seeing them in both primary and secondary markets. Ignore them at your peril.

Let’s take a quick look at the history of these markets.


Primary Market:

The primary dealers were historically a big part of treasury auctions. The allocated treasuries were mostly used to:

  • Cover ‘when issued’ (WI) forwards that they have sold to the buy-side (for speculation and fees)

  • Inventory bonds, selling them later in the secondary market to make a profit

  • Hold bonds at the Fed to meet banking reserve requirements (this was minimal pre-2008 but has since reached $2.4T)


In a nutshell, the primary market was chugging along with few changes, but unrelated external factors (i.e. the 2008 crisis) made the ride a lot bumpier. Then, there were the second order consequences of the crisis (regulation, risk controls); specifically, some enterprising hedge funds moved into the business of making money in the primary markets.

*Justice Department probes banks for rigging treasury market (June 2015), pension funds sue primary dealers for rigging treasury market (Nov 2017)

*Justice Department probes banks for rigging treasury market (June 2015), pension funds sue primary dealers for rigging treasury market (Nov 2017)

Secondary Market:

The secondary market has evolved differently, as it’s split between inter-dealer (D2D) and dealer-to-client (D2C) markets. Both markets trade roughly $270B worth of treasuries every day, mostly in on-the-run securities.

Inter-Dealer Broker market (IDB)

The inter-dealer market became orderbook-driven in the early 2000’s; since then, it has functioned as a mature, efficient market with tight spreads. Inter-Dealer Broker (IDB) platforms started off as purely inter-dealer, but they soon invited Principal Trading Firms (PTFs, such as Getco) to join, which now provide the bulk of the liquidity.


This market is fairly mature and the expected changes are minimal, but the following could happen in the near future:

  • A clearing mandate – this would remove the biggest hurdle for the buy-side on IDB but there might be side effects associated with any regulation.   

  • The FICC makes it cheap and easy for non-FICC members to clear there – PTFs won’t mind being FICC members; other buy-side firms might follow.

  • Further evolution of microstructure: bilateral streams, dark-pools, mid-point matching – There’s been lots of talk but no concrete developments yet.

Dealer to Client (D2C)

The dealer-to-client market, on the other hand, has been struggling with 3rd-world problems. Much of the trading still happens on voice or ‘Bloomberg IB’ chat. Electronification has only occurred on the operational side. Price discovery is still ‘RFQ’-based (i.e. you submit a ‘Request For Quote’ to a handful of dealers disclosing your name, trade size, and direction in many cases). The D2C electronic market is split between Tradeweb and Bloomberg. While one can argue that this is good enough or even desirable for the D2C market, it’s hard to ignore the efficiencies that the orderbook has brought to D2D Treasuries and other markets (like futures).


This market is undergoing real change. The largest has been CME’s acquisition of BrokerTec (NEX). To put it bluntly, this is a big deal for the future of the dealer-to-client treasury market. CME has been doing all they can to monetize their new asset, including making real-time market data available in Bloomberg. Rumblings on the Street suggest CME is encouraging large funds to move onto BrokerTec—a move that would have been unheard of only a year ago. It seems like CME is following the same strategy that NASDAQ tried with its eSpeed acquisition back in 2013: can they pull it off, or will CME-BrokerTec end just as badly?


While comparisons to eSpeed make sense on the surface, there are some key differences this time that make CME more likely to succeed in fully monetizing BrokerTec.

  1. CME dominates the treasury futures market and is the biggest clearing-house in the US. CME has a lot more clout to change the market than NASDAQ did back in 2013.

  2.  Unlike many years ago, PTFs are now the primary source of liquidity in the IDB market rather than dealers. With PTFs in the picture, a dealer boycott would lack teeth and end up hurting dealers more than the rest of the market.

  3. CME can incentivize its futures trading clients to move onto BrokerTec by offering margin benefits between cash and futures.

  4. There is no other option for dealers. Five years ago, eSpeed and BrokerTec each had roughly 50% market share, so dealers were able to play favorites. Today, BrokerTec owns 72% of the market and eSpeed is just 16% (some reports put it at as low as 11%).

What would it take to have the Buy-Side on BrokerTec (or any IDB/ECN)?


CME’s acquisition of BrokerTec is good news for buy-side firms hoping to gain transparent pricing and increased liquidity on BrokerTec, but some major challenges remain.

The largest is clearing: FICC membership or bilateral credit arrangements with all counterparties are required for true market players.

Today, there is a fairly complex clearing process between FICC members and PTFs (or between PTFs themselves). Essentially, BrokerTec takes on short-term credit risk and goes through a series of Prime Brokers and Custodians (as clearing, settlement, and netting agents) in order to make it all work. It is both operationally and financially inefficient, leaving lots of intraday margin netting opportunities on the table. Using TRACE data, the Fed estimates the below participant breakdown for Treasury trades done electronically on IDBs:


BrokerTec could theoretically extend that clearing methodology to the broader buy-side. Unfortunately, while that may work for very large quant funds, scaling that practice to smaller firms seems impractical. There has also been talk of clearing treasuries (see here), but that too seems unlikely in the short term due to the current administration’s opposition to regulation.


Are we there yet?

To us, this feels similar to the FX market of the early 2000s.


We think it’s likely that D2C treasuries will evolve in the same manner: more protocols and venues in a fully diversified market compared to today’s RFQ-centric model. The following large changes are poised to happen in the near future:

  • Getting the buy-side onto IDB/ECN:

o   Dealers could start to provide PB services, giving the buy-side the option to either use PB credit on IDBs or to continue trading bilaterally with dealers.

o   CME could start clearing cash treasuries, bringing a large number of D2C clients onto BrokerTec.


  • More ECNs, protocols, streams, algos

o   Bilateral streams may become a norm soon.

o   As mentioned before, a few primary dealers are offering streaming prices and algos.

o   Independent broker dealers have algos that are getting traction (Quantitative Brokers, GX2).

o   FENICS UST has a unique model for the buy-side to execute block trades at spread to the wholesale (IDB) market.

Taken together, these are strong indicators that the treasury market is becoming friendlier to the buy-side. While many buy-side funds will likely let it settle down before they alter their trading strategies, others are going to jump in and start making money, or save large sums on transaction costs.


Interest Rate Swap Primer: Origins and Overview

Interest rate swaps are an essential tool for interest rate risk management and speculation. With over 300 trillion dollars in outstanding notional and more than 10 trillion dollars traded each week, it is a staple of institutional investing: multi-national corporations, municipalities, sell-side firms, investment managers, and central banks all rely on interest rate swaps to manage interest rate risk.

Total Weekly Transaction Volume Clearing.png
Total Weekly Transaction Volume Type.png

*Source: US Commodity Futures Trading Commission’s ‘Weekly Swaps Report’ (

**Type of swaps depicted: Fixed-Float Swap; FRA (Forward Rate Agreement); OIS (Overnight Index Swap); Other ( Basis, Cap/Floor, Debt Option, Exotic, Fixed-Fixed, Inflation, and Swaption)



With IRS such an established player in the world of finance, it’s easy to forget how young these financial products really are. In order to understand the future direction of the market, it’s important to understand how the swaps market evolved into the behemoth we know today.

Swaps trading became an institutional instrument starting in 1981. At that time, the United States had an interest rate of 17%, West Germany’s rate was 11%, and Switzerland had a rate of 8%. The World Bank had reached its borrowing limits on German marks and Swiss Francs. IBM, on the other hand, had large reserves of Swiss francs and German marks, with debt payments owed in both currencies. The two parties realized they could reach a deal to leave both parties better off; together with Salomon Brothers, they engineered a specially tailored contract. Thus, the first swap contract was born.

In the early days immediately following the landmark IBM-World Bank deal, swaps were almost entirely used for hedging rather than speculation, being an attractive option for large companies seeking to protect themselves from rate hikes. As such, the swaps market was entirely made up of bilateral contracts, each with its own custom terms and conditions.

However, in the 1990’s swaps began to become increasingly popular with those looking to speculate on interest rates. Suddenly, IRS was no longer just for corporations hoping to hedge rate risk—it became a bona fide trading product, with volumes exploding through the end of the 20th century into the early aughts. While trading dwindled in the aftermath of the Global Financial Crisis, we’ve seen a steady recovery in IRS trading volumes in the post-crisis markets after 2013.

Interest Rate Derivatives.png

Source: ISDA SwapsInfo


As the IRS market grew, the products on offer became increasingly sophisticated. Instead of custom-tailored contracts for each set of counterparties, vanilla swaps emerged as a feasible standardized interest rate swap that helped to commoditize the market. This continues today with increasingly-popular MAC swaps, and the standards that have emerged for benchmarking (like the 3-month LIBOR in the US or the 6-month in Europe). At the same time, financial institutions began to customize swaps contracts in every way imaginable, creating cross-currency basis swaps, overnight index swaps, and FRAs.


Some of the popular ways swaps contracts are customized


Regulators caught on to the explosion of the swaps market and have followed suit, putting in place stringent requirements that have made swaps trading a major operational undertaking. Limited regulation started in the late 1980’s when it became clear that these wholly unregulated products carried significant risk. Most famously, this happened in the UK, where a small municipal government lost millions of pounds betting on declining interest rates. Since then, regulators have remained in lockstep with traders’ increasingly exotic swaps contracts. Consequently, IRS regulations are robust, most recently boosted by a bevy of requirements under Dodd-Frank including clearing, SEF execution, and minimum sizing standards.

With its size and regulatory oversight, it’s easy to view the swaps market similarly to any other financial marketplace. However, there are some major differences--the first is the asymmetrical standardization of swaps trading between the dealer-to-dealer market and the dealer-to-client market. Why is the market set up to favor dealers, and how will this imbalance change in the future?